The aircraft was observed to taxi from its hangar to the southern end of the airstrip, turn onto the runway and start to taxi north. After taxiing a short distance, the engine power increased and the aircraft took off to the north, becoming airborne about halfway along the strip. At a height of about 150 feet, above the northern end of the strip, the engine was heard to stop and start again. The aircraft turned left with the engine running intermittently. Soon after, it was seen in a steep nose down attitude which continued until it descended into trees.

An investigation failed to determine the reason for the intermittent operation of the engine. No defects were found with the aircraft which were likely to have contributed to the accident.

The left turn was consistent with the pilot attempting to turn back to the airfield for a landing, after the engine malfunctioned at a critical stage of flight over heavily timbered terrain, with no suitable areas for a forced landing. During the turn the aircraft apparently stalled and entered a steep nose down attitude at a height which was insufficient for the pilot to regain control.

Significant factors

The following factors were considered relevant to the development of this accident:

1. The engine malfunctioned at a critical stage of flight.

2. The aircraft was over terrain unsuitable for a safe forced landing.

3. The aircraft stalled during an attempted turnback with insufficient altitude for recovery.

On 1 July 1992, SAAB SF-340A, VH-EKT, was engaged on a scheduled passenger service from Melbourne, Victoria to Devonport, Tasmania. During the flight, the crew experienced difficulty in controlling the right propeller RPM. When the aircraft landed at Devonport, directional control was lost.

The aircraft departed the runway and ran through a ditch in soft, muddy ground. The aircraft sustained substantial damage but there were no injuries to passengers or crew. The investigation revealed that a severe asymmetric thrust condition developed after landing when reverse thrust was selected but the right propeller remained at a positive blade angle.

The report concludes that the right propeller control unit was defective, due to internal oil leakage across the feathering solenoid valve. As a result, the propeller failed to respond normally to pilot control input.

The pilot departed Moorabbin with one passenger on board and flew to the designated Moorabbin aerobatic training area with the probable intention of practising aerobatic manoeuvres. The pilot was approved to perform aerobatic manoeuvres down to 500 feet. The passenger was an experienced flying instructor who also held an approval for low level aerobatics down to 500 feet, but he was not experienced on the Pitts S2A aircraft.

Only one known witness saw the aircraft in the last few seconds before impact. Her attention was attracted to the aircraft by its loud engine noise. She briefly observed the aircraft performing an aerobatic manoeuvre while descending towards the ground at an angle of about 45 degrees. The aircraft was travelling in a westerly direction when it disappeared behind a row of Cyprus trees and impacted the ground about 300 metres north-west of her location. Subsequently the witness was shown a video of Pitts aircraft performing various aerobatic manoeuvres. The manoeuvre she identified as what she saw VH-SZA perform was a descending snap roll to the left.

Airframe and engine damage indicated that the aircraft was rotating to the left when it impacted the ground in a steep nose down attitude. Propeller damage indicated that the engine was at a low power setting at impact. No fault has been detected with the engine.

The right rudder cable attachment to the rudder horn was found to have pulled out of the Nicopress copper sleeve/swage thereby disconnecting the right rudder cable from the rudder. A similar disconnect was found with the left rudder cable at the pilot-in-command's left pedal. It was determined that the correct copper sleeves were used on the correct one eighth inch diameter rudder cables but that the swaging had been carried out with the Nicropress jaws that were appropriate for a five thirty second inch cable. Four Nicopress copper sleeves were under swaged. Specialist examination subsequently concluded that the improperly constructed rudder cables probably failed at impact and not in flight.

The reason why the two pilots failed to recover the aircraft from an aerobatic manoeuvre could not be determined.

Significant Factor

The following factor was considered relevant to the development of the accident:

1. The aircraft was engaged in low level aerobatics and struck the ground for undetermined reasons.

The pilot indicated prior to departure that his intention was to track from Launceston directly across Bass Strait to Tyabb, rather than via the Bass Strait Islands. The aircraft departed Launceston with an air traffic control clearance appropriate to that intention.

After being airborne for about one and a half hours, the aircraft approached over Cape Barren Island, although the pilot had not indicated that he was diverting from the planned track. Two very low circuits were flown in the vicinity of the airstrip and while still flying low, the pilot asked the front seat passenger to tell him what he could see below the aircraft.

The survivor, who was sitting in the right rear seat, gained the impression that the pilot could not see and that he was unwell, although apparently not in pain. He did not respond when she spoke to him. She undid her seat belt, leant over the pilot and asked if the passengers could do anything to assist. The pilot did not reply and his eyes appeared glazed. She shook him, but almost immediately the aircraft descended steeply to the ground.

From the damage sustained by the aircraft and the lack of any significant ground slide, it was assessed that the aircraft had stalled from probably no higher than 150 feet and had impacted the ground in a steep nose down and left wing low attitude. Both wing flaps were found to be extended, suggesting that the aircraft was in a landing configuration.

Medical tests and examination were unable to substantiate any cardiac or cerebral event which might have accounted for the pilot's incapacitation. There are indications, however, which suggest that the incapacitation was as a result of an insulin related condition. The pilot had successfully passed his last Civil Aviation medical examination in December 1991.

Significant Factors

The following factor was considered relevant to the development of the accident:

1. The pilot apparently suffered a subtle form of incapacitation which progressively reduced his ability to control the aircraft.

On 5 June 1992, the pilot received a one hour checkout in the aircraft from an instructor. On 12 June the pilot flew the aircraft from Coldstream to Lilydale, where his passengers were waiting. While the pilot was checking the weather forecasts an instructor commented that they indicated that the weather conditions were unfavourable. However, the pilot decided to commence the flight and assess the actual enroute weather conditions.

Melbourne Air Traffic Control Radar recorded the track and groundspeed of the aircraft from Lilydale to the area of the accident. The aircraft had passed close to Mangalore Aerodrome, where the cloud base was about 1,500 ft above ground level (agl). It then deviated from the planned track and subsequently penetrated an active military restricted area. Soldiers on the range heard the aircraft approaching and ceased all firing. The aircraft was observed to be flying at about 300 ft agl and clear of cloud. There was no rain in the immediate vicinity and the visibility was reported to be good.

After flying through this and an adjacent restricted area, the track of the aircraft became erratic. During the final minutes of the flight, the recorded groundspeeds varied from 148 kts to 75 kts. The last speed recorded was 138 kts. No radio transmissions were recorded from the aircraft.

Witnesses near the accident site observed the aircraft flying at about 300 ft and reported that the engine noise was particularly loud. Weather conditions were overcast, with the cloud base at about 500 ft and patches of rain and drizzle in the area but not near the accident site. The aircraft was then seen to bank progressively to the left and descend. Impact marks indicated that as the aircraft passed about 140 degrees of bank angle, it clipped the tops of gum trees and then struck the ground. The impact marks and degree of disintegration indicated that the aircraft impacted the ground at high speed.

A detailed examination of the wreckage did not reveal any fault or malfunction with the structure, engine and associated systems of the aircraft which might have led to the accident. The landing gear and flaps were retracted and there was ample fuel on board the aircraft. The weight and centre of gravity were estimated to have been within the prescribed limits.

The area in which the accident occurred was open and relatively flat, with several sites suitable for precautionary or forced landings. Tests conducted in the area indicated that radar coverage existed down to about 250 ft above ground level, with radio communications possible almost to ground level.

The postmortem examination revealed that the pilot had coronary artery disease, but it was not possible to determine whether this had led to an in-flight incapacitation. In addition, the pilot had a history of hypertension, which was controlled by medication. He was also on a weight loss program, which involved taking fenfluramine as an appetite suppressant. Common adverse reactions from fenfluramine include mild sedation, lethargy and giddiness.

The investigation was unable to determine the precise circumstances leading to the apparent navigation difficulties experienced by the pilot, although the weather conditions may have contributed. The behaviour of the aircraft and pilot in the vicinity of the accident site, erratic airspeed, excessive bank angle and failure to take advantage of suitable landing areas, may have been related to the physical problems suffered by the pilot. The onset of any such problems may have been too sudden to permit the pilot to conduct an emergency landing.

Significant Factors

1. Weather conditions were marginal for visual flight.

2. The pilot was probably uncertain of his geographical position.

3. It is possible that the pilot became incapacitated in flight

4. Loss of control occurred at a low height above the ground.

The aircraft was on a scheduled domestic passenger service flight from Karratha to Perth at Flight Level 310 (31,000 ft). As the aircraft entered cloud white diverting around a large thunderstorm, there was a sudden and significant rise in the outside air temperature. A short time later, all four engines progressively lost power and the aircraft was unable to maintain altitude.

During the next 17 minutes, numerous attempts to restore engine power were made without success until, approaching 10,000 ft altitude, normal engine operation was regained. The aircraft diverted to Meekatharra where a normal landing was completed. The investigation determined that during high altitude cruise, the aircraft entered an area of moist air significantly warmer than the surrounding air.

This resulted in a need to select engine and airframe anti-ice which in turn placed high bleed air demand on the engines. Under these conditions the fuel control units were unable to schedule sufficient fuel to the engines, thereby causing them to lose power, a phenomenon known as 'roll-back',

FACTUAL INFORMATION

History of the Flight

At 0700 CST on 23 July 1992, two foreign national pilots, who were undergoing airline transport pilot licence training acquired the keys and documentation for TB20 aircraft VH-LQB in preparation for a daytime instrument flight rules (IFR) navigational exercise.

The aircraft departed at about 0930. After carrying out practice instrument approaches at several aerodromes along the route, they arrived over Ceduna at 1330 where an NDB approach was made before landing. While positioning the aircraft for refuelling a scraping sound was heard from the front of the engine. After the aircraft was refuelled and the engine restarted, the same noise was heard. An examination revealed that the starter motor was remaining engaged. The pilots contacted the operator and requested assistance. The operator despatched a maintenance engineer who replaced the starter motor.

The aircraft commenced taxiing at 1854. The pilot established radio contact with Adelaide Flight Service Unit (FSU) advising that he was taxiing for Parafield, IFR, and would be using runway 11.

A Cessna Conquest, VH-ANJ, commenced taxiing at 1859, with its pilot being unable to see VH-LQB take-off. He reported taxiing to Adelaide FSU and advised that his departure would be from runway 29. Approaching the threshold of runway 11 he noticed the navigation lights of VH-LQB as the aircraft climbed, and estimated the aircraft to have been at 100 to 200 ft above ground level. He then commenced to back track along the runway for a departure to the west.

Witnesses, located to the north of the Flinders Highway, noticed the lights of an aircraft departing from Ceduna Airport, tracking in an easterly direction. Although several kilometres away from the aircrafts track, they believed it was lower than normal, having watched many aircraft depart, including three or four that day. The aircraft lights then descended towards the ground and an orange flame appeared, followed by the sound of an explosion.

VH-ANJ informed Adelaide FSU that they had observed VH-LQB depart, and then saw a fireball appear in that direction. Adelaide FSUs were unable to contact VH-LQB. VH-ANJ was requested to locate the fire when airborne and direct emergency services to it. The fire was located approximately 5 km from the airport, 3 km to the right of runway 11 centre line. It was subsequently identified as the wreckage of VH-LQB.

Injuries to persons

Both pilots of VH-LQB were fatally injured.

Damage to aircraft

The aircraft was destroyed by impact forces and post-impact fire.

Other damage

There was no other damage.

Personnel information

The pilot-in-command held a current Private Pilot Licence with a Class 1 medical certificate, and a Command Instrument Rating for single-engine aeroplanes. He was qualified to fly the TB20 aircraft and had accumulated 61 hours on the aircraft type. His total flying experience was 189 hours, which included 26 hours of instrument flight time, 26 hours of simulated instrument flight time, and 11 hours of night flying experience. He last flew at night on 17 June 1992.

The other pilot held a current Private Pilot Licence with a Class 1 medical certificate, and a Command Instrument Rating for single-engine aircraft.  He was qualified to fly the TB 20 aircraft and had accumulated 65 hours on the aircraft type. His total flying experience was 192 hours which included 38 hours of instrument flight time, and 15 hours of night flying experience. He last flew at night on 14 July 1992.

Aircraft weight and balance

The weight and centre of gravity were within limits at the time of the accident.

Meteorological information

The Ceduna weather was fine with 1 octa of strato-cumulus cloud at 3500 ft. The wind was from 170 to 180 degrees at less than 5 kts. It was reported to be a very dark night with no discernible horizon. Moon rise was not until approximately 0100.

Wreckage and impact information

Ground impact marks indicated that the aircraft had been on a heading of 150 degrees in a shallow dive, with a bank angle of 15 degrees to the right, at an estimated speed of approximately 140 kt when it impacted the ground in an area of open pasture. The right wing separated from the fuselage. The aircraft continued in the southerly direction, breaking up as it went, with the cabin consumed by fire. Propeller contact with the ground caused a torsional failure of the crankshaft at the propeller attachment flange. The propeller travelled in a direction of 210 degrees for 65 m, leaving a trail of slash marks in the ground as it continued to rotate. The engine left wing and other components separated during the impact sequence. The engine bounced away on a curved track to the right for a distance of 110 m in a direction of 240 degrees from the main wreckage trail. The remains of the fuselage, consisting of the fire gutted cabin and tailplane, came to rest 166 m from the initial point of impact.

Fire damage was confined to the cabin area, which was totally consumed, except for the two instrument panels and co-pilot's seat, which were thrown clear.

No evidence was found to suggest a pre-existing defect in the aircraft structure or control systems prior to the accident. The propeller damage, condition of the engine and its components, and specialist analysis of exhaust pipe temperatures at impact, indicated that the engine was delivering power at impact. The temperature of the exhaust pipes was determined to be approximately 500 to 600 degrees Celsius at the time of impact. The vacuum pump frangible coupling, made from a ductile polymeric material, had failed under torsional load. Laboratory examination determined that there was no rotational damage to the failed coupling faces which indicated that the engine ceased operating immediately following the coupling failure.

Medical and pathological information

There was no evidence that either pilot had any medical or psychological condition which might have contributed to the accident.

Fire

A fierce fire, fuelled by aviation gasoline from the ruptured fuel tanks, engulfed the aircraft's cabin, reducing to ash the area between the engine firewall and the tailplane. The remainder of the wreckage suffered little or no fire damage.

Tests and research

Flight profile

A similar aircraft was used to attempt to replicate the accident flight profile from various altitudes. Six flights were conducted. The first flight followed the Ceduna to Adelaide track on climb to a point abeam the accident site. Normal climb power settings and speeds were used. The next four flights used a 10 to 15 degree angle of bank to the right, with the aircraft being held in a dive so as to arrive at the accident site without changing the climb power or trim settings during the descent. The final flight was conducted with power off and the controls left unattended from the top of climb.

The investigation was not able to establish the actual flight profile of the aircraft. However, from the flights it was considered possible that the aircraft may have descended from approximately 500 ft at an indicated airspeed of 140 kts.

The effects of somatogravic illusion were considered. Acceleration after take-off, into a dark horizonless sky accentuates the illusion giving the impression of a steep nose-up attitude. This illusion may cause a pilot to assume that the aircraft attitude is to nose high, and to respond by lowering the aircraft nose. However, studies have shown that this illusion is only likely to have an effect during the take-off rotation and initial climb phase of flight.

Attitude indicator

Tests were later conducted with a similar type of aircraft to establish the time for the attitude indicator to topple following a vacuum source failure.

This was simulated by ground running the engine at 2000 RPM for 5 minutes to ensure the gyro was fully erect, then shutting it down. This test, with a vacuum indication of 4.8 inches of mercury was conducted several times resulting in an indication of a bank to the left.

Gyro instruments

Detailed inspection of the gyro instruments failed to reveal any defects which may have prevented their normal operation.

Additional information

Following replacement of the starter motor the battery was found to be flat requiring the use of an external power source for starting. It was after last light by the time the repairs had been completed, and the pilot who had flown the aircraft to Ceduna carried out the engine run following the starter motor replacement. He then remained in the left pilot seat during the preparations for departure, as it was considered easier to leave the engine running than jump start it again, so by default he became the pilot-in-command for the return flight.

While talking to the pilots prior to departure the instructor who accompanied the engineer to Ceduna did not consider giving them a night operations briefing but was satisfied that they were capable of conducting the flight.

Flight and duty times

On the day of the accident the pilots had planned to complete the exercise during daylight hours. They had commenced duty at 0700, prepared and submitted their flight plan, checked the aircraft, then departed at approximately 0930.  One pilot being in command for the Adelaide to Ceduna sector, the other for the return sector to Adelaide. The flight to Ceduna took approximately 4 hours.

Training

Prior to commencement of night flying training, this operator ensured that all student pilots received classroom instruction in night flying techniques. In the syllabus of training the operator also covered the subjects of somatogravic and other illusions which can effect pilots under limited visual conditions.

The initial night flying training was carried out at Parafield, and students were subsequently taken to country airports, away from the lights of built-up areas, so they could experience dark night operations.

ANALYSIS

Following an apparent normal take-off the aircraft then descended back into the ground in what appeared to be controlled flight.

It is unlikely that the aircraft was lower than 500 ft above ground level when it commenced the descent. At this height the effects of acceleration and therefore somatogravic illusion would not be experienced by the pilot.

After becoming airborne the pilot may have become involved in recording the departure time, making a departure call on the CTAF, and changing to the Adelaide FSU frequency to advise them of the departure and next estimate, while failing to give sufficient attention to flying the aircraft.

At departure the pilots had been on duty for 12 hours, and it is possible that fatigue may have contributed to the accident.

The reason for the accident could not be established.

The aircraft was operating on to a 617 m grass strip with a fence at each end. The orientation of the strip was north/south and threshold markers were positioned on either side of the strip 79 m in from the fence at the northern end. Some 40 m beyond the southern end of the strip was a group of trees 14 m high.

The weather at the time of the accident was fine with good visibility. The Bureau of Meteorology estimated the surface wind to have been from the north-west at 12 kt. There was no windsock at the landing strip.

Observers saw the aircraft, apparently operating normally, fly towards the landing area and make an approach to land to the south. One witness said the aircraft touched down at the threshold markers. A second witness reported that the aircraft touched down about 200 m in from the fence, bounced to a height of 3-4 ft, then touched down again about 25 m further on. The witnesses then saw the aircraft continue along the strip at speed, with a significant level of engine power applied. Approaching the southern end of the strip, the aircraft swerved right, probably to avoid the group of trees beyond the end of the strip. The witnesses heard an increase in engine noise and saw the aircraft become airborne. The landing gear contacted the boundary fence, and the aircraft continued across a road, through a line of bushes, and struck the side of a house. An intense fire broke out, destroying the aircraft and much of the building. The pilot escaped from the wreckage, but the passenger did not.

The first marks identified on the airstrip surface were main wheel tyre tracks which commenced about 300 m from the threshold. These marks indicated that skidding had occurred from 350 m from the threshold and continued for about 50 m. From there the aircraft ran straight ahead for about 170 m before veering to the right by approximately 12-15 degrees. The tyre tracks continued to within about 20 m of the boundary fence.

Inspection of the wreckage did not reveal evidence of any defects that might have contributed to the accident. The landing gear was down, and the wing flaps were extended 17 degrees.

The Aircraft Flight Manual landing chart does not provide for landing distance calculation where the tailwind component exceeds five knots. The Beechcraft A36 Pilot's Operating Handbook indicated that, with a 10 kt tailwind and when approaching over a 50 ft obstacle, a landing distance of about 700 metres would be required. However, when approaching over a 20 ft obstacle, this distance is reduced to about 500 m.

The available evidence indicates that the pilot conducted an approach to land in a manner which was not appropriate in the prevailing conditions. The pilot misjudged the approach and touched down well into the strip. The decision to go-around was delayed until too late to safely complete the manoeuvre.

The pilot declined to make himself available for interview during the investigation.

Significant Factors

The following factors were considered relevant to the development of the accident:

1. The pilot attempted a landing with a significant tail wind component.
2. The landing approach was misjudged.
3. An attempt to go around was made with insufficient airstrip remaining.

The Boeing 747 (QFA 27) was cleared for take-off from runway 16 on a Mudgee One Standard Instrument Departure (SID). This SID procedure required the aircraft to track via the 163 radial of the Sydney VOR (omni) to 3,000 ft and then turn left to track back towards the airfield, passing overhead the Sydney VOR prior to setting course in a north-westerly direction.

The VOR navigational aid is located on Sydney Airport and the SID required the B747 to be at 5,000 ft or higher before passing the VOR. Sydney Air Traffic Control had cleared the aircraft to climb to its initial cruising level, flight level (FL) 310. The Bell 206 (VH-BHU) planned to climb overhead Sydney Airport to FL 125 for a photographic operation.

The Sydney Aerodrome Controller (ADC) cleared the helicopter to climb to FL 125 within the lateral confines of an area between Qantas Maintenance facilities, located on the airport and the Sydney Hilton Hotel which is about 0.5 km west of the airport's northern boundary. Due to suppression of radar returns within 3.5 km of the radar head, which is also located on the airport, the flight path of the helicopter was not detected on any ATC radar screen.